python pandas - identify row of rolling maximum value from panel?

Question

Does Panel have anything like pan.idxmax(axis='items') which could return the row # or index for the largest value as in the question linked below which piRSquared answered perfectly?

Link to original question

Answer 1

consider the pd.Panel p

dfs = dict(
    one=pd.DataFrame(np.random.randint(1, 10, (5, 5))),
    two=pd.DataFrame(np.random.randint(1, 10, (5, 5))),
    three=pd.DataFrame(np.random.randint(1, 10, (5, 5))),
)

p = pd.Panel(dfs)
p.to_frame().unstack()

---

the pd.Panel object doesn't have an idxmax method. But the underlying numpy array does have the argmax method. I've built this custom function to leverage that.

def idxmax(pn, axis=0):
    indices = pd.Series(['items', 'major_axis', 'minor_axis'])
    idx, col = indices.drop(axis)
    return pd.DataFrame(pn.values.argmax(axis),
                        pn.__getattribute__(idx),
                        pn.__getattribute__(col))

---

usage

idxmax(p, 0)

idxmax(p, 1)

idxmax(p, 2)